Memory and Music

By Alex Russell - A lot happens when we remember. Networks of neurons firing throughout the brain let us see, hear, smell, touch and maybe even taste something that happened in the past. An interdisciplinary research project at UC Davis is using music to access memories that on most days seem buried under years of living—and forgetting.

Innovative memory research has wide-ranging implications

  • This interdisciplinary research project was funded in 2015 by the Institute for Social Sciences, and includes PI Petr Janata and collaborators Raul Aranovich, Duncan Temple Lang and Arne Ekstrom.
  • Music is a powerful way to evoke distant memories, and could provide a way to better understand how memories are encoded in the brain and later recalled.
  • The project will produce a database of music-evoked memories generated by visitors to the project's website, which will be a useful tool for researchers as well as families with loved ones in palliative care.
  • Learn more about the project at Petr Janata's lab website.

The research team, led by Professor of Psychology Petr Janata, is working to connect a person’s known memories evoked by clips of music to regions in the brain that light up on scans when the music is heard again. The results will expand our knowledge on how the brain encodes memories. It could also provide a way to improve quality of life for those suffering debilitating conditions including Alzheimer’s.

“This technique is really at the fore of neuroimaging research right now,” says Janata, who won a Guggenheim Fellowship in 2010 for his work on music and autobiographical memory. “There are various types of experiments that show people large libraries of images and then try to reconstruct the images based on their brain activity. That’s similar to this idea but no one has done it with music.”

Studying Music and Memory

In January 2015, Janata and his team won two years of seed funding from the Institute for Social Sciences to expand this work. The collaboration includes Associate Professor of Linguistics Raul Aranovich, Professor of Statistics and director of the Data Science Initiative Duncan Temple Lang and Associate Professor of Psychology Arne Ekstrom. 

One part of the project is to expand their Web platform for crowdsourcing written memories. The website plays a 30-second clip of music and then asks questions like, “On a scale of one to five, how strong does a memory experience does it evoke?” The questions are all pre-programmed based on how a visitor responds.

If a visitor reports that a music clip evokes a strong memory, they are prompted to write that memory out. Then they are asked to play another clip.

The MEAM Website

In a preliminary version of the Music-Evoked Autobiographical Memory (MEAM) website, visitors can hear clips of music and submit memories. Those who submit more than 200 memories--the raw material for building a neurobiography--might be selected to take part in fMRI scanning as part of the study.

Try it out.

Temple Lang will help by processing the memory reports to yield general concepts. Aranovich will then identify key words and concepts from the memory reports and distill them down to about half a dozen nouns, verbs and adjectives. Those keywords will help the team to build a graph that shows how the same concepts might also appear in the brain from song to song during a scan.

“All of this is language,” says linguist Raul Aranovich, who studies language structure and theory. “Events in the text could become part of a web of meaning,” he says.

A handful of visitors to the website have already submitted hundreds of these music-evoked memories. The team plans to invite some of these “super-users” to undergo a brain scan while they listen to those same clips. That way, the researchers may be able to identify overlaps between the language-based concepts in the written memories and areas in the brain that light up on the brain scan when they hear the music.

“You get away from these overly rehearsed memories and instead you have a way of getting at these remote, forgotten events,” says Janata. 

How the Brain Remembers

Music, says Janata, could prove to be a powerful avenue to learn more about how memories are stored in the brain. It affords a much broader set of memory retrieval cues than images. This is important, because part of the problem with studying memory is that there is no such thing as a perfect memory. Something else always gets in the way.

“You can never encode the memory very well in the first place," says Ekstrom, "and if you don’t encode something very well in the first place you’re not going to retrieve it very well.”

In his Spatial Cognition Lab at UC Davis, Ekstrom studies spatial navigation and virtual reality. He and Janata have another study that places participants into virtual environments to navigate while hearing music in different locations. This way they can compare music with location for how well each lets people retrieve memories.

“In general,” says Ekstrom, “people tend to remember the most when they are given both music and location as memory cues.”

Brain Cells and Mapping

Arne Ekstrom’s lab at the UC Davis Center for Neuroscience investigates how spatial memories happen in the brain, how they are recorded and how different types of memories are processed. In a 2003 study, Ekstrom and his team asked patients who were undergoing seizure monitoring to navigate virtual environments on a laptop computer.

While the patients completed these tasks, Ekstrom's team could record the activity of individual cells because the patients had sensors implanted in the brain as part of their monitoring for seizures. These recordings showed that individual cells in the temporal lobes of the brain fired at specific locations in the virtual environment, even regardless of the location’s orientation on the screen.

The classic view of memory, he says, is like a digital camera. Presumably, all you have to do is scroll through what has been recorded. A different way to think about memory is when someone tries to remember a trip to Chicago but ends up thinking about Boston instead. 

In reality, he says, research is showing that memory only gathers enough information to provide a general gist of what happened. Recalling accurate, detailed memories is very hard to do. “People can retrieve them,” says Ekstrom, “but it’s not as common as we think.”

A song, says Ekstrom, might bring back even distant memories from summer camp 20 years ago. In a similar way, a visit to your hometown might bring otherwise forgotten memories flooding back.

Scanning for Memories

To capture those memories as they take place in the brain, the team will use functional magnetic resonance imaging (fMRI) to record brain activity as a person listens to clips of music for which they have already provided written memories. For the scan, a person will lie down in a giant tube where a strong magnetic field and radio waves create detailed images that represent changes in blood flow in the brain.

On the monitor screen during the scan, changes in the brain's blood flow lights up tiny 3D cubes of brain area called “voxels,” which are measured in millimeters rather than at scales not visible to the eye, like individual cells.

“It’s difficult to directly record from the brain,” says Ekstrom. “We can only stick electrodes into the brain in human patients in situations where people are undergoing monitoring for surgical planning, unless there is a clinical mandate.”

Janata is hoping to pinpoint the collections of voxels throughout the brain that show more activity when a participant listens to a clip of music that sparks a powerful memory. By knowing beforehand the content of those memories, it might be possible to create a map--a “neurobiography”--of those memories.

Janata says it’s unlikely that memories with shared emotional content or even similar experiences occupy the same space in different people’s brains. Beyond the major structures in the brain that everyone has, the uniqueness of individual life experience connects networks of brain cells differently.

“The likelihood you’re going to converge on the same wiring diagram for two people who were at the same Grateful Dead show is highly unlikely,” he says.

Future Opportunities and Palliative Care

In addition to ISS seed funding for this project, Janata and Temple Lang were also selected as ISS Fellows to offer a special graduate student course in fall 2015. Graduate students who enroll in this proseminar will learn advanced tools for analyzing and processing some of the data that come out of the team’s memory research.

So far Janata’s team has complete data, brain scans and all, on one person who wrote over 33,000 words in memory reports based on over 200 pieces of music. Janata recorded six hours of fMRI data as she listened to 157 memory-evoking pieces of music and viewed 66 personal photographs from eight distinct periods of her life. Just from her participation, Janata says he can already start to see how memories are represented in the brain.

Another goal in this line of research is to create a database that allows for the mapping of the semantic content and interrelationships of millions of memories and the clips of music that evoked them. The database could be of immense value for researchers across the social sciences, says Aranovich.

“Having this multidimensional structure could help us understand human nature,” he says. "It also shows how language relates to other sensory issues in cognition.”

Janata says that an even more important outcome from this project would be to build a cost-effective mechanism for improving psychological health and palliative care. The online library of memories could be a way for families to build musical memory scrapbooks. These would increase social bonding which helps with palliative care. 

“You can imagine the people going through cancer and chemotherapy, people who need their life outlook boosted," says Janata. "This can be a valuable tool."

Janata says that a lot of research shows that nostalgia boosts optimism and helps maintain an overall positive mood. This is critical for Alzheimer’s patients, for whom memories are especially dear, but also for patients facing any number of other debilitating conditions.

Ekstrom says he is excited to work on a project that takes such a radically different approach to studying memory.

“Almost no one in psychology studies music,” says Ekstrom. “In scientific disciplines you have the tried and true way of doing things. Janata is saying, ‘Let’s clear the table and try something different.’ I’m really happy to be part of that.”

Filed under: